Yes, your tree is balanced as per the definition of an AVL tree. Typically, we are interested in balanced search trees. Thus, you should consider the binary search tree (BST) properties as well, along with the height balancing condition.
Now, to address your confusion, try to prove the following lemma:
For an AVL tree with $n$ nodes whose balance factor lies in $\{−1, 0,+1\}$, the height $h$ lies in the interval $$\log_2(n + 1) − 1 ≤ h < \log_\phi(n + 2) + b$$ where $\phi$ is the golden ratio and $b \approx −1.3277$.
As long as the height $h$ is within a logarithmic bound of the number of nodes $n$ in the tree, we are in a good place.
If new inserion/deletion is infrequent in your AVL tree compared to the number of search queries, it is better to restructure the nodes so that the height minimizes.
People have also experimented with various kinds of balancing schemes to better suit their requirements. One such realization is described here: https://en.wikipedia.org/wiki/Weight-balanced_tree.